There is an discussion about the same at Sun.
http://forum.java.sun.com/thread.jsp?thread=386190&forum=48&message=1660462
I would use a datasource or pool to ensure a flow in the application unless you're
very sure,
that your connection implementation allows concurrent statements to be executed and
not just
executes them one at a time.
The link above reads :
>>For example, two Statements on the same Connection can be executed
>>concurrently and their ResultSets can be processed concurrently (from the
>>perspective of the
developer). Some drivers will provide this full
>>concurrency. Others may execute one statement and wait until it completes before
>>sending the next.
Maby you should look at Jarkata' pool or look at some datasources.
Few connections are good because fewer connections uses less memory and less cpu and
therefore
allows faster executions.
But fewer connections also, depending on the complexity of the requests and size of
the results,
makes "faster waits" for those threads not in possession of a connection :@)
Regards
Kleth
------------------------------------------------------------------------------------------
Kristian A. Leth
Maersk Data Transport/Architects & Specialists.
------------------------------------------------------------------------------------------
Hans Wichman
<[EMAIL PROTECTED]> To: [EMAIL PROTECTED]
cc:
05-08-2003 21:29 Subject: database connection
Please respond
to "Tomcat Users
List"
Hi,
I am evaluating a jsp application for a customer, which provides an
interface to a database through some helperclasses.
Each helper class manages the retrieval (and updates) of related
underlying tables, and there are about 10 helperclasses.
In the header (header.jsp) the jsp calls an initialize method on a
Configuration object, checks if it is already initialized
(this process is not synchronized), and if not, creates one connection,
creates all the 10 helper classes as static instances passing this
one-and-only connection.
In the rest of the application a collection of tags call the static methods
(unsynchronized) on the Configuration object, such as
getDatabaseHelper1 (be it with a more descriptive name) and performs
operations on the underlying database through this class.
All the methods in the Helper classes are unsynchronized as well, but they
are standalone (utility functions, I don't know a better way to describe it).
My question is, except from any race conditions in the helper classes
themselves, is it a problem that all those helper classes share only one
connection and access it unsynchronized. If so, why is it a problem if
unsynchronized communication occurs on a database connection? Or are all
the underlying methods on the connection object already synchronized and is
there no need to worry?
Some pointers to why and how would be appreciated as well. And I am not
only looking to why connection pooling would be good, but also why one
connection would be bad ;-)
kind regards
Hans
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
